aboutsummaryrefslogtreecommitdiff
path: root/vendor/gmp-6.3.0/mpn/powerpc64/mode64/sqr_basecase.asm
blob: e76bb8878d81527c55928fb94d47930ee4d450e1 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
dnl  PowerPC-64 mpn_sqr_basecase.

dnl  Contributed to the GNU project by Torbjorn Granlund.

dnl  Copyright 1999-2001, 2003-2006, 2008, 2010, 2011 Free Software Foundation,
dnl  Inc.

dnl  This file is part of the GNU MP Library.
dnl
dnl  The GNU MP Library is free software; you can redistribute it and/or modify
dnl  it under the terms of either:
dnl
dnl    * the GNU Lesser General Public License as published by the Free
dnl      Software Foundation; either version 3 of the License, or (at your
dnl      option) any later version.
dnl
dnl  or
dnl
dnl    * the GNU General Public License as published by the Free Software
dnl      Foundation; either version 2 of the License, or (at your option) any
dnl      later version.
dnl
dnl  or both in parallel, as here.
dnl
dnl  The GNU MP Library is distributed in the hope that it will be useful, but
dnl  WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
dnl  or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
dnl  for more details.
dnl
dnl  You should have received copies of the GNU General Public License and the
dnl  GNU Lesser General Public License along with the GNU MP Library.  If not,
dnl  see https://www.gnu.org/licenses/.

include(`../config.m4')

C                  cycles/limb
C POWER3/PPC630         6-18
C POWER4/PPC970          8
C POWER5                 8
C POWER6                16.25
C POWER7                 3.77

C NOTES
C  * This is very crude, cleanup!
C  * Try to reduce the number of needed live registers.
C  * Rewrite for POWER6 to use 8 consecutive muls, not 2 groups of 4.  The
C    cost will be more live registers.
C  * Rewrite for POWER7 to use addmul_2 building blocks; this will reduce code
C    size a lot and speed things up perhaps 25%.
C  * Use computed goto in order to compress the code.
C  * Implement a larger final corner.
C  * Schedule callee-saves register saves into other insns.  This could save
C    about 5 cycles/call.  (We cannot analogously optimise the restores, since
C    the sqr_diag_addlsh1 loop has no wind-down code as currently written.)
C  * Should the alternating std/adde sequences be split?  Some pipelines handle
C    adde poorly, and might sequentialise all these instructions.
C  * The sqr_diag_addlsh1 loop was written for POWER6 and its preferences for
C    adjacent integer multiply insns.  Except for the multiply insns, the code
C    was not carefully optimised for POWER6 or any other CPU.
C  * Perform cross-jumping in sqr_diag_addlsh1's feed-in code, into the loop.

C INPUT PARAMETERS
define(`rp', `r3')
define(`up', `r4')
define(`n',  `r5')

define(`rp_outer', `r25')
define(`up_outer', `r21')
define(`rp_saved', `r22')
define(`up_saved', `r23')
define(`n_saved',  `r24')

ASM_START()
PROLOGUE(mpn_sqr_basecase)
	cmpdi	cr0, n, 2
	bge	cr0, L(ge2)
	ld	r5, 0(up)	C n = 1
	nop
	mulld	r8, r5, r5	C weight 0
	mulhdu	r9, r5, r5	C weight 1
	std	r8, 0(rp)
	std	r9, 8(rp)
	blr
	ALIGN(16)
L(ge2):	bgt	cr0, L(gt2)
	ld	r0, 0(up)	C n = 2
	nop
	mulld	r8, r0, r0	C u0 * u0
	mulhdu	r9, r0, r0	C u0 * u0
	ld	r6, 8(up)
	mulld	r10, r6, r6	C u1 * u1
	mulhdu	r11, r6, r6	C u1 * u1
	mulld	r4, r6, r0	C u1 * u0
	mulhdu	r5, r6, r0	C u1 * u0
	addc	r4, r4, r4
	adde	r5, r5, r5
	addze	r11, r11
	addc	r9, r9, r4
	adde	r10, r10, r5
	addze	r11, r11
	std	r8, 0(rp)
	std	r9, 8(rp)
	std	r10, 16(rp)
	std	r11, 24(rp)
	blr

	ALIGN(16)
L(gt2):	std	r31,  -8(r1)
	std	r30, -16(r1)
	std	r29, -24(r1)
	std	r28, -32(r1)
	std	r27, -40(r1)
	std	r26, -48(r1)
	std	r25, -56(r1)
	std	r24, -64(r1)
	std	r23, -72(r1)
	std	r22, -80(r1)
	std	r21, -88(r1)

	mr	rp_saved, rp
	mr	up_saved, up
	mr	n_saved, n
	mr	rp_outer, rp
	mr	up_outer, up

	rldicl.	r0, n, 0,62	C r0 = n & 3, set cr0
	cmpdi	cr6, r0, 2
	addic	r7, n, 2	C compute count...
	srdi	r7, r7, 2	C ...for ctr
	mtctr	r7		C copy count into ctr
	beq-	cr0, L(b0)
	blt-	cr6, L(b1)
	beq-	cr6, L(b2)

L(b3):	ld	r6, 0(up)
	ld	r9, 8(up)
	ld	r27, 16(up)
	addi	up, up, 24
	li	r12, 0		C carry limb
	bdz	L(em3)

	ALIGN(16)
L(tm3):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	ld	r9, 0(up)
	ld	r27, 8(up)
	adde	r0, r0, r12
	adde	r7, r7, r26
	mulld	r26, r9, r6
	mulhdu	r10, r9, r6
	mulld	r11, r27, r6
	mulhdu	r12, r27, r6
	ld	r9, 16(up)
	ld	r27, 24(up)
	std	r0, 8(rp)
	adde	r26, r26, r8
	std	r7, 16(rp)
	adde	r11, r11, r10
	std	r26, 24(rp)
	addi	up, up, 32
	std	r11, 32(rp)
	addi	rp, rp, 32
	bdnz	L(tm3)

L(em3):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	adde	r0, r0, r12
	adde	r7, r7, r26
	std	r0, 8(rp)
	std	r7, 16(rp)
	addze	r8, r8
	std	r8, 24(rp)
	addi	n, n, 2
	b	L(outer_loop)

L(b0):	ld	r6, 0(up)
	ld	r27, 8(up)
	mulld	r7, r27, r6
	mulhdu	r12, r27, r6
	std	r7, 8(rp)
	addi	rp, rp, 8
	ld	r9, 16(up)
	ld	r27, 24(up)
	addi	up, up, 32
	bdz	L(em0)

	ALIGN(16)
L(tm0):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	ld	r9, 0(up)
	ld	r27, 8(up)
	adde	r0, r0, r12
	adde	r7, r7, r26
	mulld	r26, r9, r6
	mulhdu	r10, r9, r6
	mulld	r11, r27, r6
	mulhdu	r12, r27, r6
	ld	r9, 16(up)
	ld	r27, 24(up)
	std	r0, 8(rp)
	adde	r26, r26, r8
	std	r7, 16(rp)
	adde	r11, r11, r10
	std	r26, 24(rp)
	addi	up, up, 32
	std	r11, 32(rp)
	addi	rp, rp, 32
	bdnz	L(tm0)

L(em0):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	adde	r0, r0, r12
	adde	r7, r7, r26
	std	r0, 8(rp)
	std	r7, 16(rp)
	addze	r8, r8
	std	r8, 24(rp)
	addi	n, n, 2
	b	L(outer_loop_ent_2)

L(b1):	ld	r6, 0(up)
	ld	r9, 8(up)
	ld	r27, 16(up)
	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r12, r27, r6
	addc	r7, r7, r26
	std	r0, 8(rp)
	std	r7, 16(rp)
	addi	rp, rp, 16
	ld	r9, 24(up)
	ld	r27, 32(up)
	addi	up, up, 40
	bdz	L(em1)

	ALIGN(16)
L(tm1):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	ld	r9, 0(up)
	ld	r27, 8(up)
	adde	r0, r0, r12
	adde	r7, r7, r26
	mulld	r26, r9, r6
	mulhdu	r10, r9, r6
	mulld	r11, r27, r6
	mulhdu	r12, r27, r6
	ld	r9, 16(up)
	ld	r27, 24(up)
	std	r0, 8(rp)
	adde	r26, r26, r8
	std	r7, 16(rp)
	adde	r11, r11, r10
	std	r26, 24(rp)
	addi	up, up, 32
	std	r11, 32(rp)
	addi	rp, rp, 32
	bdnz	L(tm1)

L(em1):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	adde	r0, r0, r12
	adde	r7, r7, r26
	std	r0, 8(rp)
	std	r7, 16(rp)
	addze	r8, r8
	std	r8, 24(rp)
	addi	n, n, 2
	b	L(outer_loop_ent_3)

L(b2):	addi	r7, r7, -1	C FIXME
	mtctr	r7		C FIXME
	ld	r6, 0(up)
	ld	r9, 8(up)
	ld	r27, 16(up)
	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	ld	r9, 24(up)
	mulld	r11, r9, r6
	mulhdu	r10, r9, r6
	addc	r7, r7, r26
	adde	r11, r11, r8
	addze	r12, r10
	std	r0, 8(rp)
	std	r7, 16(rp)
	std	r11, 24(rp)
	addi	rp, rp, 24
	ld	r9, 32(up)
	ld	r27, 40(up)
	addi	up, up, 48
	bdz	L(em2)

	ALIGN(16)
L(tm2):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	ld	r9, 0(up)
	ld	r27, 8(up)
	adde	r0, r0, r12
	adde	r7, r7, r26
	mulld	r26, r9, r6
	mulhdu	r10, r9, r6
	mulld	r11, r27, r6
	mulhdu	r12, r27, r6
	ld	r9, 16(up)
	ld	r27, 24(up)
	std	r0, 8(rp)
	adde	r26, r26, r8
	std	r7, 16(rp)
	adde	r11, r11, r10
	std	r26, 24(rp)
	addi	up, up, 32
	std	r11, 32(rp)
	addi	rp, rp, 32
	bdnz	L(tm2)

L(em2):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	adde	r0, r0, r12
	adde	r7, r7, r26
	std	r0, 8(rp)
	std	r7, 16(rp)
	addze	r8, r8
	std	r8, 24(rp)
	addi	n, n, 2
	b	L(outer_loop_ent_0)


L(outer_loop):
	addi	n, n, -1
	addi	up_outer, up_outer, 8
	addi	rp_outer, rp_outer, 16

	mr	up, up_outer
	addi	rp, rp_outer, 8

	srdi	r0, n, 2
	mtctr	r0

	bdz	L(outer_end)

	ld	r6, 0(up)
	ld	r9, 8(up)
	ld	r27, 16(up)
	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	ld	r9, 24(up)
	ld	r28, 0(rp)
	ld	r29, 8(rp)
	ld	r30, 16(rp)
	mulld	r11, r9, r6
	mulhdu	r10, r9, r6
	addc	r7, r7, r26
	adde	r11, r11, r8
	addze	r12, r10
	addc	r0, r0, r28
	std	r0, 0(rp)
	adde	r7, r7, r29
	std	r7, 8(rp)
	adde	r11, r11, r30
	std	r11, 16(rp)
	addi	rp, rp, 24
	ld	r9, 32(up)
	ld	r27, 40(up)
	addi	up, up, 48
	bdz	L(ea1)

	ALIGN(16)
L(ta1):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6	C 9
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6	C 27
	ld	r9, 0(up)
	ld	r28, 0(rp)
	ld	r27, 8(up)
	ld	r29, 8(rp)
	adde	r0, r0, r12	C 0 12
	adde	r7, r7, r26	C 5 7
	mulld	r26, r9, r6
	mulhdu	r10, r9, r6	C 9
	mulld	r11, r27, r6
	mulhdu	r12, r27, r6	C 27
	ld	r9, 16(up)
	ld	r30, 16(rp)
	ld	r27, 24(up)
	ld	r31, 24(rp)
	adde	r26, r26, r8	C 8 5
	adde	r11, r11, r10	C 10 11
	addze	r12, r12	C 12
	addc	r0, r0, r28	C 0 28
	std	r0, 0(rp)	C 0
	adde	r7, r7, r29	C 7 29
	std	r7, 8(rp)	C 7
	adde	r26, r26, r30	C 5 30
	std	r26, 16(rp)	C 5
	adde	r11, r11, r31	C 11 31
	std	r11, 24(rp)	C 11
	addi	up, up, 32
	addi	rp, rp, 32
	bdnz	L(ta1)

L(ea1):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	ld	r28, 0(rp)
	ld	r29, 8(rp)
	adde	r0, r0, r12
	adde	r7, r7, r26
	addze	r8, r8
	addc	r0, r0, r28
	std	r0, 0(rp)
	adde	r7, r7, r29
	std	r7, 8(rp)
	addze	r8, r8
	std	r8, 16(rp)

L(outer_loop_ent_0):
	addi	n, n, -1
	addi	up_outer, up_outer, 8
	addi	rp_outer, rp_outer, 16

	mr	up, up_outer
	addi	rp, rp_outer, 8

	srdi	r0, n, 2
	mtctr	r0

	ld	r6, 0(up)
	ld	r9, 8(up)
	ld	r27, 16(up)
	ld	r28, 0(rp)
	ld	r29, 8(rp)
	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	addc	r0, r0, r28
	adde	r7, r7, r26
	addze	r12, r8
	std	r0, 0(rp)
	adde	r7, r7, r29
	std	r7, 8(rp)
	addi	rp, rp, 16
	ld	r9, 24(up)
	ld	r27, 32(up)
	addi	up, up, 40
	bdz	L(ea0)

	ALIGN(16)
L(ta0):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6	C 9
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6	C 27
	ld	r9, 0(up)
	ld	r28, 0(rp)
	ld	r27, 8(up)
	ld	r29, 8(rp)
	adde	r0, r0, r12	C 0 12
	adde	r7, r7, r26	C 5 7
	mulld	r26, r9, r6
	mulhdu	r10, r9, r6	C 9
	mulld	r11, r27, r6
	mulhdu	r12, r27, r6	C 27
	ld	r9, 16(up)
	ld	r30, 16(rp)
	ld	r27, 24(up)
	ld	r31, 24(rp)
	adde	r26, r26, r8	C 8 5
	adde	r11, r11, r10	C 10 11
	addze	r12, r12	C 12
	addc	r0, r0, r28	C 0 28
	std	r0, 0(rp)	C 0
	adde	r7, r7, r29	C 7 29
	std	r7, 8(rp)	C 7
	adde	r26, r26, r30	C 5 30
	std	r26, 16(rp)	C 5
	adde	r11, r11, r31	C 11 31
	std	r11, 24(rp)	C 11
	addi	up, up, 32
	addi	rp, rp, 32
	bdnz	L(ta0)

L(ea0):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	ld	r28, 0(rp)
	ld	r29, 8(rp)
	adde	r0, r0, r12
	adde	r7, r7, r26
	addze	r8, r8
	addc	r0, r0, r28
	std	r0, 0(rp)
	adde	r7, r7, r29
	std	r7, 8(rp)
	addze	r8, r8
	std	r8, 16(rp)

L(outer_loop_ent_3):
	addi	n, n, -1
	addi	up_outer, up_outer, 8
	addi	rp_outer, rp_outer, 16

	mr	up, up_outer
	addi	rp, rp_outer, 8

	srdi	r0, n, 2
	mtctr	r0

	ld	r6, 0(up)
	ld	r9, 8(up)
	ld	r28, 0(rp)
	mulld	r0, r9, r6
	mulhdu	r12, r9, r6
	addc	r0, r0, r28
	std	r0, 0(rp)
	addi	rp, rp, 8
	ld	r9, 16(up)
	ld	r27, 24(up)
	addi	up, up, 32
	bdz	L(ea3)

	ALIGN(16)
L(ta3):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6	C 9
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6	C 27
	ld	r9, 0(up)
	ld	r28, 0(rp)
	ld	r27, 8(up)
	ld	r29, 8(rp)
	adde	r0, r0, r12	C 0 12
	adde	r7, r7, r26	C 5 7
	mulld	r26, r9, r6
	mulhdu	r10, r9, r6	C 9
	mulld	r11, r27, r6
	mulhdu	r12, r27, r6	C 27
	ld	r9, 16(up)
	ld	r30, 16(rp)
	ld	r27, 24(up)
	ld	r31, 24(rp)
	adde	r26, r26, r8	C 8 5
	adde	r11, r11, r10	C 10 11
	addze	r12, r12	C 12
	addc	r0, r0, r28	C 0 28
	std	r0, 0(rp)	C 0
	adde	r7, r7, r29	C 7 29
	std	r7, 8(rp)	C 7
	adde	r26, r26, r30	C 5 30
	std	r26, 16(rp)	C 5
	adde	r11, r11, r31	C 11 31
	std	r11, 24(rp)	C 11
	addi	up, up, 32
	addi	rp, rp, 32
	bdnz	L(ta3)

L(ea3):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	ld	r28, 0(rp)
	ld	r29, 8(rp)
	adde	r0, r0, r12
	adde	r7, r7, r26
	addze	r8, r8
	addc	r0, r0, r28
	std	r0, 0(rp)
	adde	r7, r7, r29
	std	r7, 8(rp)
	addze	r8, r8
	std	r8, 16(rp)


L(outer_loop_ent_2):
	addi	n, n, -1
	addi	up_outer, up_outer, 8
	addi	rp_outer, rp_outer, 16

	mr	up, up_outer
	addi	rp, rp_outer, 8

	srdi	r0, n, 2
	mtctr	r0

	addic	r0, r0, 0
	li	r12, 0		C cy_limb = 0
	ld	r6, 0(up)
	ld	r9, 8(up)
	ld	r27, 16(up)
	bdz	L(ea2)
	addi	up, up, 24

	ALIGN(16)
L(ta2):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6	C 9
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6	C 27
	ld	r9, 0(up)
	ld	r28, 0(rp)
	ld	r27, 8(up)
	ld	r29, 8(rp)
	adde	r0, r0, r12	C 0 12
	adde	r7, r7, r26	C 5 7
	mulld	r26, r9, r6
	mulhdu	r10, r9, r6	C 9
	mulld	r11, r27, r6
	mulhdu	r12, r27, r6	C 27
	ld	r9, 16(up)
	ld	r30, 16(rp)
	ld	r27, 24(up)
	ld	r31, 24(rp)
	adde	r26, r26, r8	C 8 5
	adde	r11, r11, r10	C 10 11
	addze	r12, r12	C 12
	addc	r0, r0, r28	C 0 28
	std	r0, 0(rp)	C 0
	adde	r7, r7, r29	C 7 29
	std	r7, 8(rp)	C 7
	adde	r26, r26, r30	C 5 30
	std	r26, 16(rp)	C 5
	adde	r11, r11, r31	C 11 31
	std	r11, 24(rp)	C 11
	addi	up, up, 32
	addi	rp, rp, 32
	bdnz	L(ta2)

L(ea2):	mulld	r0, r9, r6
	mulhdu	r26, r9, r6
	mulld	r7, r27, r6
	mulhdu	r8, r27, r6
	ld	r28, 0(rp)
	ld	r29, 8(rp)
	adde	r0, r0, r12
	adde	r7, r7, r26
	addze	r8, r8
	addc	r0, r0, r28
	std	r0, 0(rp)
	adde	r7, r7, r29
	std	r7, 8(rp)
	addze	r8, r8
	std	r8, 16(rp)

	b	L(outer_loop)

L(outer_end):
	ld	r6, 0(up)
	ld	r9, 8(up)
	ld	r11, 0(rp)
	mulld	r0, r9, r6
	mulhdu	r8, r9, r6
	addc	r0, r0, r11
	std	r0, 0(rp)
	addze	r8, r8
	std	r8, 8(rp)

define(`rp',  `rp_saved')
define(`up',  `r5')
define(`n',   `r6')
define(`climb',	`r0')

	addi	r4, rp_saved, 8
	mr	r5, up_saved
	mr	r6, n_saved

	rldicl.	r0, n, 0,62		C r0 = n & 3, set cr0
	cmpdi	cr6, r0, 2
	addi	n, n, 2			C compute count...
	srdi	n, n, 2			C ...for ctr
	mtctr	n			C put loop count into ctr
	beq	cr0, L(xb0)
	blt	cr6, L(xb1)
	beq	cr6, L(xb2)

L(xb3):	ld	r6,   0(up)
	ld	r7,   8(up)
	ld	r12, 16(up)
	addi	up, up, 24
	mulld	r24, r6, r6
	mulhdu	r25, r6, r6
	mulld	r26, r7, r7
	mulhdu	r27, r7, r7
	mulld	r28, r12, r12
	mulhdu	r29, r12, r12
	ld	r10,  8(rp)
	ld	r11, 16(rp)
	ld	r6,  24(rp)
	ld	r7,  32(rp)
	addc	r10, r10, r10
	adde	r11, r11, r11
	adde	r6, r6, r6
	adde	r7, r7, r7
	addze	climb, r29
	addc	r10, r10, r25
	adde	r11, r11, r26
	adde	r6, r6, r27
	adde	r7, r7, r28
	std	r24,  0(rp)
	std	r10,  8(rp)
	std	r11, 16(rp)
	std	r6,  24(rp)
	std	r7,  32(rp)
	addi	rp, rp, 40
	bdnz	L(top)
	b	L(end)

L(xb2):	ld	r6,  0(up)
	ld	r7,  8(up)
	addi	up, up, 16
	mulld	r24, r6, r6
	mulhdu	r25, r6, r6
	mulld	r26, r7, r7
	mulhdu	r27, r7, r7
	ld	r10,  8(rp)
	ld	r11, 16(rp)
	addc	r10, r10, r10
	adde	r11, r11, r11
	addze	climb, r27
	addc	r10, r10, r25
	adde	r11, r11, r26
	std	r24,  0(rp)
	std	r10,  8(rp)
	std	r11, 16(rp)
	addi	rp, rp, 24
	bdnz	L(top)
	b	L(end)

L(xb0):	ld	r6,   0(up)
	ld	r7,   8(up)
	ld	r12, 16(up)
	ld	r23, 24(up)
	addi	up, up, 32
	mulld	r24, r6, r6
	mulhdu	r25, r6, r6
	mulld	r26, r7, r7
	mulhdu	r27, r7, r7
	mulld	r28, r12, r12
	mulhdu	r29, r12, r12
	mulld	r30, r23, r23
	mulhdu	r31, r23, r23
	ld	r10,  8(rp)
	ld	r11, 16(rp)
	ld	r6,  24(rp)
	ld	r7,  32(rp)
	ld	r12, 40(rp)
	ld	r23, 48(rp)
	addc	r10, r10, r10
	adde	r11, r11, r11
	adde	r6, r6, r6
	adde	r7, r7, r7
	adde	r12, r12, r12
	adde	r23, r23, r23
	addze	climb, r31
	std	r24,  0(rp)
	addc	r10, r10, r25
	std	r10,  8(rp)
	adde	r11, r11, r26
	std	r11, 16(rp)
	adde	r6, r6, r27
	std	r6,  24(rp)
	adde	r7, r7, r28
	std	r7,  32(rp)
	adde	r12, r12, r29
	std	r12, 40(rp)
	adde	r23, r23, r30
	std	r23, 48(rp)
	addi	rp, rp, 56
	bdnz	L(top)
	b	L(end)

L(xb1):	ld	r6,  0(up)
	addi	up, up, 8
	mulld	r24, r6, r6
	mulhdu	climb, r6, r6
	std	r24, 0(rp)
	addic	rp, rp, 8		C clear carry as side-effect

	ALIGN(32)
L(top):	ld	r6,   0(up)
	ld	r7,   8(up)
	ld	r12, 16(up)
	ld	r23, 24(up)
	addi	up, up, 32
	mulld	r24, r6, r6
	mulhdu	r25, r6, r6
	mulld	r26, r7, r7
	mulhdu	r27, r7, r7
	mulld	r28, r12, r12
	mulhdu	r29, r12, r12
	mulld	r30, r23, r23
	mulhdu	r31, r23, r23
	ld	r8,   0(rp)
	ld	r9,   8(rp)
	adde	r8, r8, r8
	adde	r9, r9, r9
	ld	r10, 16(rp)
	ld	r11, 24(rp)
	adde	r10, r10, r10
	adde	r11, r11, r11
	ld	r6,  32(rp)
	ld	r7,  40(rp)
	adde	r6, r6, r6
	adde	r7, r7, r7
	ld	r12, 48(rp)
	ld	r23, 56(rp)
	adde	r12, r12, r12
	adde	r23, r23, r23
	addze	r31, r31
	addc	r8, r8, climb
	std	r8,   0(rp)
	adde	r9, r9, r24
	std	r9,   8(rp)
	adde	r10, r10, r25
	std	r10, 16(rp)
	adde	r11, r11, r26
	std	r11, 24(rp)
	adde	r6, r6, r27
	std	r6,  32(rp)
	adde	r7, r7, r28
	std	r7,  40(rp)
	adde	r12, r12, r29
	std	r12, 48(rp)
	adde	r23, r23, r30
	std	r23, 56(rp)
	mr	climb, r31
	addi	rp, rp, 64
	bdnz	L(top)

L(end):	addze	climb, climb
	std	climb,  0(rp)

	ld	r31,  -8(r1)
	ld	r30, -16(r1)
	ld	r29, -24(r1)
	ld	r28, -32(r1)
	ld	r27, -40(r1)
	ld	r26, -48(r1)
	ld	r25, -56(r1)
	ld	r24, -64(r1)
	ld	r23, -72(r1)
	ld	r22, -80(r1)
	ld	r21, -88(r1)
	blr
EPILOGUE()